Automatic Face Expression Detection

Authors

  • D. Dervinis Šiauliai University

Abstract

The method for detecting face expression in the static image is analyzed. Method is able to detect one from five typical face expressions (angry, happy, neutral, sadness, surprise). Algorithm uses eighteen points in eyebrows, eyes, lips. There were used 325 separate human expressions in experiment. Statistically were calculated changes of points coordinates in different expressions between neutrally and final expression images and defined important changes of points pairs. In order to detect face expression using this method, neutral expression must be taken in to account. In detection process were used „City Block“ and neural multilayer perceptron network methods. Total accuracy of the method is 87 %. Ill. 9, bibl. 9 (in Lithuanian; summaries in Lithuanian, English and Russian).

Downloads

Published

2005-10-20

How to Cite

Dervinis, D. (2005). Automatic Face Expression Detection. Elektronika Ir Elektrotechnika, 64(8), 79-83. Retrieved from https://eejournal.ktu.lt/index.php/elt/article/view/10514

Issue

Section

T 115 MEDICINE TECHNOLOGY